Using Buggify

Deterministic fault injection uses a stable seed to activate chaos at specific call sites, ensuring race conditions and rare timing bugs are reproducible across simulation runs. This allows you to consistently trigger, study, and fix hidden failure paths.

The best way to apply chaos testing to external calls and state transitions is to wrap them with buggify! or buggify_with_prob! macros, using strategic placement guidance to inject failures at resource boundaries and timeouts.

You track coverage when injecting faults by using sometimes_assert macros to verify that buggify triggers activate during the simulation. This provides observability and preserves safety while exercising rare error paths.

Yes, you can reproduce a chaos testing scenario across different simulation providers by using a fixed deterministic seed. This stable seed ensures deterministic randomness and consistent buggify behavior regardless of the underlying provider.

Yes, you need a deterministic seed to reproduce bugs found during chaos testing. The seed-based strategy activates chaos at specific call sites, which is required to maintain reproducibility and consistently study the resulting bug path.

Your buggify fault injections may not be triggering because the probability is too low or the placement is incorrect. Adjust probabilities using buggify_with_prob! and consult the provided TROUBLESHOOTING and PLACEMENT-GUIDE materials to correct the chaos site.